Cooperative computing for space applications using advanced instrument controllers

This study examines a distributed execution system composed of C51-based miniaturized computing devices called advanced instrument controllers (AICs). Developed by Phillips Laboratories and taking up roughly the space of a postage stamp, AICs combine a digital signal processor with 128K of EEPROM, 128K of RAM, 32 analog input channels, 8 analog channels, and 6 communication ports in a radiation-hardened package. The distributed execution system uses AICs with a task distribution system called MOM, and a planning tool called generational scheduling. MOM is a fault-tolerant support system executing across distributed, heterogeneous network of processing elements which has the ability to diagnose faults, identify recovery of nodes, and rollback lost jobs. Generational scheduling (GS) is simplistic yet effective scheme for dynamically scheduling dependent tasks in a near-optimal manner on available resources. By using AICs, MOM, and GS a self-organizing and self-compensating execution environment is obtained